Chest CT plays an important role in the diagnostic and prognostic assessment of COVID-19 pneumonia. The CT findings of COVID-19 pneumonia are complex and, include ground-glass nodules or ground-glass opacities, lung consolidation, thickening of the interlobular septum, "fibrosis-like" strips, air bronchogram signs, and thickening of blood vessels. Among these findings, "fibrosis-like" strips are found in some COVID-19 patients in both lungs on the first chest CT examination. At present, the exact imaging and pathological mechanism of "fibrosis-like" strips are still unclear. Many radiological experts in China believe that this finding is indicative of pulmonary fibrosis . However, owing to the lack of systematic imaging studies and adequate pathological evidence, this conclusion is still controversial. Clinically, the prognosis of pulmonary fibrosis is poor, and the identification of pulmonary fibrosis would help clinicians determine whether the patient needs anti-fibrosis treatment. Therefore, the qualitative diagnosis of this sign has critical clinical significance. Lung biopsy is the gold standard for the diagnosis of pulmonary fibrosis. However, as an invasive procedure, lung biopsy not only increases the risk to patients, but also increases the risk of infection by medical staff. Therefore, it is difficult to obtain pathological specimens through lung biopsy in clinical practice.
Similar to COVID-19 pneumonia, SARS-CoV pneumonia is often combined with pulmonary fibrosis. It has been reported that the incidence of pulmonary fibrosis in SARS-CoV is as high as 25.9-62% [3-4]. An autopsy study  regarding SARS-related deaths demonstrated that pulmonary fibrosis appeared as early as 38 days after infection with the SARS virus. Wang et al.  also found that strip-shaped fibrosis in patients with SARS first appeared 30-40 days after admission by follow-up CT. Lu et al.  performed a CT follow-up study and observed that in 10 SARS patients with varying degrees of pulmonary fibrosis, the fibrosis was absorbed within 6 months, and pulmonary fibrosis in two patients lasted for 6 months and 12 months. Chen et al.  also reported that 17.1% (19/111) of SARS patients still had residual pulmonary fibrosis 18 months after discharge. These studies suggest that pulmonary fibrosis in SARS patients has characteristics of late appearance and slow absorption.
A total of 36 COVID-19 patients with "fibrosis-like" strips were included in the present study. The initial chest CT scan on admission showed "fibrosis-like" strips in both lungs (Fig. 1a, 1e, 1i, 1m, 2a, 2b, 2e, 2f) in the majority of patients (29/36, 80.6%), with an average appearance time of 2 days (1-3 days). Seven patients (7/36, 19.4%) had “fibrosis-like” strips in both lungs on the second CT scan, with an average appearance time of 10 days (4-17 days). The average appearance time of "fibrosis-like" strips for all 36 patients was 2.7 days (1-17 days). All patients had different degrees of absorption of "fibrosis-like" strips after active intervention, and 41.7% of the patients showed complete absorption of "fibrosis-like" strips in the bilateral lungs (median 37 days, range 19-124 days). A total of 22.2% of the patients showed complete absorption of "fibrosis-like" strips in the unilateral lung (median 36 days, range 24-59 days). Additionally, 36.1% of the patients, showed a significant reduction of "fibrosis-like" strips in the bilateral lungs. Compared with previously published data regarding SARS-related pulmonary fibrosis [3-4, 5-6, 8], "fibrosis-like" strips in COVID-19 pneumonia seem to appear earlier and be absorbed faster. Therefore, based on the imaging manifestations and outcome characteristics of "fibrosis-like" strips in the patients in the present study, we conclude that "fibrosis-like" strips in COVID-19 patients may not be pulmonary fibrosis.
On the first CT scan performed on day 1 at hospital in patients 6, "fibrosis-like" strips appeared in the bilateral lungs, and bronchia contained in the strips were observed (Fig. 2i). Moreover, the following three CT re-examinations (Fig. 2j-2l) showed that the number of strips with bronchia gradually decreased. Additionally, we used the bronchium contained in the "fibrosis-like" strip as the targeted bronchus for curved reformatting (Fig. 2m-2p), and observed a small nodular high-density shadow in the bronchium. Four consecutive CT scans (Fig. 2m-2p) showed that small nodular high-density shadows not only gradually decreased in number, but also changed in shape and position. Therefore, we speculate that the small nodular high-density shadows in the bronchium may be mucus embolus, while the strips may be pulmonary tissue atelectasis caused by the mucus embolus of the bronchium. The changes in the number, shape, and location of the small nodular high-density shadows in the bronchium suggest that they may indicate the process of mucus emboli being discharged from bronchia. At present, several pathological studies on COVID-19 pneumonia have also shown that mucus-like materials adhere to the walls of bronchia and/or mucus emboli in bronchia [9-13].
Fig. 3a-v shows the evolution process of the "fibrosis-like" strips in another case of COVID-19 pneumonia, and the spatial relationships between the bronchia and strips were displayed visually on curved reformatted images (Fig. 3s-3v). On the first CT examination of the patient, bronchia contained in the "fibrosis-like" strips were observed in the bilateral lungs (Fig. 3a, 3g, 3m), which was considered inflammation in the lower lobes of both lungs with localized pulmonary atelectasis; the third CT scan showed decreased consolidation but increased strips in the bilateral lungs (Fig. 3c, 3i, 3o), suggesting reduced inflammation, but progressed pulmonary atelectasis. Subsequently, the fourth and fifth CT scans showed that the number of "fibrosis-like" strips containing bronchia (Fig. 3d, 3j, 3p) gradually decreased or even evolved into solid "fibrosis-like" strip (Fig. 3e, 3k, 3q), which may be associated with pulmonary atelectasis tissues gradually returning to normal when the inflammation was gradually absorbed after treatment. Therefore, the bronchia contained in the pulmonary atelectasis tissues also became gradually smooth and returned to normal; thus, the bronchia in the "fibrosis-like" strips gradually became invisible. The "fibrosis-like" strips were completely absorbed 43 days after onset (Fig. 3f, 3l, 3r). This typical case clearly demonstrates the evolution of the spatial relationships between "fibrosis-like" strips and bronchia over time, indicating that "fibrosis-like" strips are pulmonary segmental atelectasis rather than pulmonary fibrosis.
Although pulmonary fibrosis was observed in COVID-19 patients in several autopsy studies [9, 11, 12], most of these death cases were related to severe and critical pneumonia, and they were associated with at least one underlying disease; therefore, the occurrence of pulmonary fibrosis may be associated with severe lung injury in these severe and critical cases.
For patients with mild and common COVID-19 pneumonia, the lung injury is relatively mild; in addition, the majority of "fibrosis-like" strips contain bronchia, and most of these lesions can be absorbed in a short period, which does not correspond to the imaging characteristics of pulmonary fibrosis.
In conclusion, "fibrosis-like" strips on chest CT scans in COVID-19 patients have the characteristics of early appearance and rapid morphological changes as well as rapid absorption. In addition, bronchia contained in "fibrosis-like" strips were observed in the majority of patients (88.9%) in this study. Therefore, we speculate that "fibrosis-like" strips may not be pulmonary fibrosis, but a sign of pulmonary segmental atelectasis. However, the number of patients presenting with this typical evolution was relatively small, and all of the included patients in this study had mild or common COVID-1. The evidence is still insufficient, and further imaging-pathological correlation studies are needed to confirm our results.
Despite its limitations, this study certainly offers valuable insights into the potential pathogenesis of "fibrosis-like" strips in COVID-19 patients.